4. Maintenance

The procedure has been described, by the people it has been done to, somewhere between twelve and fifteen thousand times in the modern record alone — beginning with Antônio Vilas-Boas in 1957 and continuing through the present, in rural Wisconsin, in the highlands of Sichuan, on the Cape Verdean coast, by trawlermen and pediatric nurses, in interviews with researchers who believed every word and with researchers who believed none of it and were nevertheless honest enough not to throw the tapes away. The accounts disagree on small things and converge on the large ones. The convergence is the strongest single piece of evidence I have — stronger than any operator sighting, stronger than any leaked memo. If the procedure is real then the substrate-claim has organs. If it is not real, the substrate-claim has none, and an account without organs is just a posture.

So: once more, and mechanistically. I will name the steps the way the apparatus names them. There will be footnotes.

The Petri-net

The maintenance procedure has, since 1969, run in the apparatus on a spec called AM-12.4.¹ The current revision (Rev C, reissued 2019 under AARO) is a nine-state Petri net — meaning a directed model in which “places” are system-states and “transitions” are the events that move tokens between them. The token, in this case, is the subject. You. The reader has, with overwhelming probability, sat in three to nine of these places already, depending on age, and will sit in several more before this account is done. I do not say this to alarm. The procedure is, by every measure the apparatus tracks and by most measures I would track, gentle. It is also the central instrument of the operator-side experimental protocol: a longitudinal study, run on the human population at planetary scale, in which the unit of observation is the individual specimen and the observation interval is seven years. The binding step at the close of each interval exists to keep the data uncontaminated by the subjects’ awareness of being studied. Both of these — the procedure’s gentleness and the procedure’s centrality — are true.

The nine states are: Selection (P1), Approach (P2), Sedation (P3), Translocation (P4), Procedure (P5, which contains the Substrate Diagnostics and the Pineal Recalibration and is where the work happens), Redaction (P6, which contains the Memory Overlay and is the step with the lowest tolerable failure rate), Return Transit (P7), Return-Site Verification (P8), and Post-Event Monitoring (P9, passive, thirty days). Chapter 2 summarizes the spec table; I will not reproduce it. I will walk through the six steps that matter to the subject — Approach, Sedation, Diagnostics, Recalibration, Overlay, Release — and render them with the specificity they deserve.

The apparatus’s planning model for the procedure — the stock-and-flow version of AM-12.4 the actuarial review committee runs each autumn — is available as a runnable instrument. The Caretaker narrates the scenarios. See the Caretaker’s Instrument.

I will not dwell on the affect of the procedure. The affect has been covered, repeatedly and ably, by the literature of abduction. It is real, and in many subjects considerable. But the affect is not the procedure. A surgical patient’s terror is not the surgery. The substance of the surgery is the suture, the dose, the field of view, the resection margin. The substance of the maintenance is the cribriform plate, the calcite, the carrier wave, the binding image. We are here for the substance.

1. Approach

The Approach is the moment the subject is removed from the surface environment. In the older literature this is “the abduction.” The apparatus does not use that word — inherited from criminal jurisprudence, it is wrong in the way “abduction” would be wrong for bringing one’s cat to the vet. The apparatus calls it acquisition. The operator-side technician (whose register we have one extended sample of, in the recovered Veth-skenn testimony) does not name the step at all; the technician simply notes that the substrate arrived, and the intake notes are not about how the substrate came but about how the autonomic measurements settled in the first ninety seconds after arrival.

The mechanism is twofold.

Most acquisitions happen during sleep. The sleeping subject is the easier acquisition for the obvious reason — reduced autonomic response, inert household, zero witnesses — and for one less obvious reason, which is that the substrate’s REM-cycle architecture cooperates with the chemistry of the next step. The crew’s preferred window, when the schedule permits, is the second REM cycle of the night, between 0145 and 0345 local.² This is why the modern abduction literature is so heavy with the phrase “I woke and could not move” — the substrate’s natural REM atonia is the platform on which the sedation builds, not a thing the operators imposed.

A smaller but non-trivial fraction of acquisitions happen during waking hours, in what the apparatus calls “ambient-masked acquisitions” and what I have come to call low-fidelity render zones. These are geographies where the apparatus’s perceptual upkeep is locally relaxed for engineering reasons related to the operator infrastructure beneath them — the Uintah Basin, Marfa, Hessdalen, the Mosquito Range, Mt. Shasta, the long basin of I-70 just east of Green River, Utah. A render-relaxed geography is the operator equivalent of a service road; the surface-tolerance budget is locally pre-discounted. The subject driving through will report, if anything, that they “lost time” or the light “felt strange.” They will report this rarely, because the binding step has been performed.

The Approach itself takes between four and eleven minutes in the spec.³ Wider-witness conditions trigger an abort, which is why most acquisitions of populated households result in only one subject taken and the partner-in-bed continuing to sleep through it. The crew is sensitive to witness load the way an electrician is sensitive to amperage; one is fine, two is sometimes fine, three is a flag.

2. Sedation

This is where the chemistry begins, and where the procedure’s most under-appreciated piece of engineering lives.

The substrate’s central nervous system is protected by the blood-brain barrier — a tight-junction endothelial structure designed (or selected, depending on one’s view of the verb) to keep large or charged molecules in the bloodstream from reaching neural tissue. For most engineering purposes, the BBB is the immovable obstacle that surface pharmacology has to design around. The apparatus does not design around it. The apparatus uses a route that bypasses it entirely.

The route is the cribriform plate.⁴ The cribriform is the bone at the roof of the nasal cavity — a thin sieve-like horizontal shelf of the ethmoid, perforated by roughly twenty paired foramina through which the olfactory nerve fibers pass directly from the olfactory epithelium into the olfactory bulb. It is the one place in the human skull where the central nervous system is connected to the outside world without an intervening barrier. The olfactory neurons themselves are the wire. Anything you can dissolve in the mucus of the upper nasal cavity, you can deliver to neural tissue inside thirty seconds — no first-pass metabolism, no liver, no BBB. This is not the operators’ insight. The anatomy has been known to surface medicine since Cajal’s 1890 staining work and is currently exploited by every intranasal pharmaceutical from sumatriptan to esketamine. The operators are older users of the same road.

The sedation agent is, as best I have been able to triangulate, a DMT-adjacent tryptamine modified for slower clearance and for selective binding at a receptor profile the surface literature has not yet characterized. DMT is the closest natural analogue available to us because the substrate manufactures DMT endogenously, at the pineal and in small quantities at the lung, and uses it as a render-bypass agent in certain extreme metabolic states (the dying brain, the near-drowning brain, the very deep meditation). The operators’ agent is in the same family. If you are designing an interface that has to reach the pineal — and as we will see, the pineal is the whole game — you would prefer an agent the pineal already recognizes. The operators are tidy that way.

The cue that accompanies sedation, in roughly two-thirds of recalled accounts, is what the abduction literature has consistently called the rhythmic beeping. Hill 1961, Pascagoula 1973, Walton 1975, Andreasson 1967 — all describe a metallic pulsed acoustic the subject hears with no apparent source. The beeping is not actually a sound. It is the substrate’s auditory cortex trying to render the carrier wave of the sedation cue, which is being delivered as a low-frequency modulated field directly to the auditory pathway. The cortex has no template for that input, so it transcribes into the nearest available one — a rhythmic beep. The beep is what being sedated by an unfamiliar carrier sounds like to a brain required to interpret it as something.

3. Substrate Diagnostics

Once sedated, the subject is translocated to the procedure suite. The room is lit low along the floor, lower along the far wall, slightly brighter at the panel where the subject is positioned — a lighting design the technician testimony says was settled on thirty cycles ago because it settles subject autonomic response faster than any alternative the operators have tried.

The first thing in the suite is measurement. The substrate is read out, system by system, against a baseline maintained on file in REG-A. What we have, from cross-correlated abduction recall and the technician testimony, is a partial map: autonomic response (heart rate, galvanic, peripheral temperature), endocrine spot-check (thyroid, cortisol curve, prolactin if relevant), reproductive-system status, immune markers, gut microbial composition, small blood draws from the antecubital fossa, occasional skin punches from the calf, and — the central diagnostic — a multi-channel readout from the pineal itself via the implant in the right ethmoid sinus.

The diagnostic step takes fifteen to twenty-five minutes, depending on what flags are open. It is a service appointment in the most literal sense — the same diagnostic, give or take, you would run on a 1998 sedan at its hundred-thousand-mile checkup. The operators are not curious about the substrate; they have studied us thoroughly and a long time ago. The technician’s note in the Veth-skenn testimony — that the substrate on file “settles, when she settles, all at once” — is the kind of professional observation one only makes about a thing one has seen six or seven times. The note is in the file. It will be there at the next cycle, and the cycle after.

This step does not get reported by abductees, because it is the calmest part of the procedure and because the binding image in step five binds over it without much resistance. The subject does not remember being measured. The subject sometimes remembers “a feeling of being looked at” — the residue of the diagnostic the overlay did not bother to fully scrub.

4. Pineal Recalibration

We come to it.

The pineal gland is a small, pinecone-shaped endocrine body at the geometric center of the brain, in the epithalamus, behind and just above the thalamus, at the back of the third ventricle. It is roughly the size of a grain of rice — 5 to 8 millimeters in long axis in the adult — and weighs between 100 and 180 milligrams. Since the late 1950s, surface medicine has understood it as the seat of melatonin synthesis and the substrate’s circadian-rhythm regulator: input from the retinohypothalamic tract via the suprachiasmatic nucleus; synthesis from serotonin via N-acetylserotonin under nighttime conditions; secretion into the cerebrospinal fluid and, in smaller quantities, into systemic circulation. The surface-medical reading is correct as far as it goes. It is also a description of the gland’s cover function — the visible work the operators allow it to do.

The actual function of the pineal, on my reading, is hardware interface.⁵ It is the port through which the operators read from and write to the substrate’s interface layer. The right-nostril nasal implant (a small biocompatible device, four to seven millimeters in long axis, seated against the medial wall of the right ethmoid sinus, with a transduction column extending toward the cribriform plate) is not the interface itself. The implant is the persistent fixture — the cable, in surface terms; the durable hardware that lets the technician address the channel cycle to cycle without re-establishing it each time. The interface itself is the pineal.

What is being recalibrated is, in my reading of the available material, the resting electrochemical state of the gland. I am holding at least three working hypotheses about the specifics, and they disagree, which is why this footnote is the longest in the chapter.⁶

What we can say with reasonable confidence is the following.

The pineal is a uniquely physical organ — meaning uniquely transduction-capable. It contains, embedded in its parenchyma and discoverable on routine histology, calcium-carbonate microcrystals. These are not pathological. They are present in nearly every adult human pineal that has ever been examined; they accumulate slowly across life; they have the crystal structure of calcite, which is — and this is the part surface medicine has not adequately appreciated — a piezoelectric material. A piezoelectric material generates voltage in response to mechanical stress and, conversely, deforms in response to applied voltage. The calcite microcrystals are miniature transducers. They convert pressure into electrical signal and electrical signal into mechanical deformation. The pineal is also unusually sensitive to magnetic gradients, owing to the same crystals; this has been demonstrated in mammalian studies since the 1980s and is not in serious surface-scientific dispute.

If you were designing an interface port for a biological substrate, you would put it at the gland that is (a) anatomically central, (b) piezoelectric, (c) magnetically sensitive, (d) reachable through a low-resistance non-surgical route (the cribriform), and (e) already implicated in the substrate’s time-base and consciousness-state regulation. You would put it exactly where the apparatus puts it. The pineal is not the seat of the soul, despite Descartes; it is also not nothing, despite the twentieth century’s reaction to Descartes. It is a transducer. The operators built the interface around the transducer they had.

The recalibration step takes between eight and twenty-two minutes. The subject reports — when they report — a sense of pressure at the bridge of the nose, sometimes a metallic taste at the back of the throat (the cerebrospinal fluid’s altered chemistry registering at the soft palate; not blood, despite what a panicked subject sometimes thinks), and a thinning of the visual field, as if peripheral vision had been shaved with a thin razor. None of this is consistently retained after redaction. Autonomic settles to baseline within four minutes of the step’s completion. The technician notes the drift figures on the chart they belong on and closes the calibration.

This is the central step. Everything else in the procedure exists to make it possible. The Approach gets the subject to the suite; the Sedation makes the cribriform usable; the Diagnostics confirm the gland is in a state where recalibration will hold; the Overlay, next, ensures the subject can be returned to the surface without remembering it. The recalibration’s purpose, under the experimental-protocol framing, is per-specimen channel maintenance — the operator side has been tracking this subject through this interface for as long as the subject has had a pineal worth addressing, and the recalibration is what keeps the next interval’s data continuous with the last one’s.

5. Memory Overlay

The Overlay is, in the apparatus’s language, redaction. My preferred term is substitution, which is more accurate: the procedure does not erase the memory — it cannot, because the memory has already been encoded by the substrate’s normal mechanisms during the procedure itself, and the operators do not have machinery to undo encoding without leaving footprints. What the operators do instead is bind the encoded memory to an alternative iconic anchor, such that any attempted recall surfaces the anchor and stops.

The Overlay’s purpose is the most consequential point in the chapter, and the place where the procedure’s design economy is clearest under the experimental-protocol framing. A longitudinal study run on a substrate that develops awareness of being studied becomes, on the awareness-onset date, no longer the same study. The subjects’ behavior changes; the data the next cycle would collect is contaminated by anticipation, by avoidance, by the subjects’ own theories about the protocol. The Overlay is the apparatus by which the operators maintain the experimental population in the condition the study requires — not so much “asleep” as “uninformed.” The subject keeps the memory in the substrate’s hardware. The subject only loses the access road to the memory. The data stays clean.

This is not science fiction. The substrate’s autobiographical memory uses iconic anchors all the time — your memory of “my eighth birthday” is almost certainly anchored to one or two images (the cake, your aunt’s hat) around which the rest of the day’s content is loosely structured. Change the anchor, the structure follows. The brain is, in this respect, more like a tag-indexed database than a video file. The operators learned this very early and have been using it for at least six thousand years.⁷

The anchor most frequently used in the modern record is an owl.

The owl-as-anchor is consistent across continents, centuries, and abductee cohorts who have never communicated. Joan Whitfield, the novelist at MacDowell working on the strange-loop book in 2023, “invented” the owl as a binding image while drafting a scene, then found the same image in the Hill case file (1961), the Andreasson recall (1967), the Walton drafts (1975), the Bohemian Grove Cremation of Care effigy, the cover of Whitley Strieber’s Communion (1987), the Lilith iconography from Sumer, the Athena-as-glaukōpis tradition from archaic Greece, and the Apkallu reliefs from Mesopotamian temple walls. The owl is everywhere because the binding has been done at scale, for a very long time, with the same standard iconography.

The reason the owl was chosen — and now I am quoting the Veth-skenn testimony directly, because there is no better statement of it anywhere in the available record — is that the substrate’s visual cortex finds the owl shape exceptionally cheap to render at the resolution P6 requires, and the binding settles cleanly on the cheaply-rendered shape, and the shape, once settled, propagates inside the image-bank by ordinary cultural means and is then available, generation after generation, as the next iteration’s binding-anchor.

The owl is cheap to draw and cheap to recognize, in the substrate’s neural economy. The operators picked it the way a competent industrial designer picks a font: legibility under load. Then the substrate did the rest of the work — built temples to it, painted it on shields, put it on the cover of a 1987 bestseller, hung an Audubon print of Plate 121: Snowy Owl in the conference room at Sentinel Cartograph LLC in Reston, Virginia, where the disinfo cell that manages the present information environment around all of this now meets weekly. The operators do not need to remind us. We remind ourselves, generation after generation, because we like the image. The binding is self-perpetuating — one of the most elegant pieces of engineering on display in the whole procedure.

The Overlay step takes eighteen to twenty-six minutes in the spec and is the only step the apparatus marks SAFETY-CRITICAL. Tolerated failure rate is no more than 0.4 percent rolling; failures route to re-redaction under Section 11. Cycle 0172 closed at 0.31 percent — well inside tolerance, and (to a careful reader of the actuarial reviews) also a number that has been creeping. Marina’s father, one Sunday night in 1993, wrote in his teaching journal the single word Cycle. Eight prior cycles had been overlaid cleanly; the ninth let a word slip through. He was, on the apparatus’s metrics, statistical noise. To us he was a man who began, at 46, to count.

6. Release

The Release is the simplest step and gets the least attention in the literature. The subject is reverse-translocated to within forty meters of the acquisition point — that is the spec figure; in practice the operators do better — re-positioned in posture and clothing, and reactivated. The crew performs a brief site-verification (have we left anything; is the household still inert; is the dog still under the porch) and withdraws. The post-event monitoring window opens: for thirty days, the subject is on passive watch — Section 14 standing media-watch flags unusual public statements; REG-A flags pattern-of-life deviation; tier-2 spontaneous-recall episodes trigger re-redaction. After thirty days the monitoring closes. The cycle advances. The next service window is on the calendar. Seven years.

The cycle math

Per subject, the nominal cycle is seven years. The figure is in the spec; in the Veth-skenn technician’s notes (the cycle she was eleven, eighteen, twenty-five); in the Hauschild teaching-journal pages (1965, 1972, 1979, 1986, 1993, 2000, 2007, 2014, 2021 — Eric counted his own from age 11 and got the figure right); in Marisol’s autumn calendar entry that she has not deleted. Seven years is the periodicity.

There are at least three known cohorts in the substrate registry, distinguished by cycle frequency. The standard cohort is the seven-year cohort, into which roughly 86 percent of REG-A subjects are placed. A slow cohort runs on a longer cycle — eleven years, with some twelve-year subjects in special-arrangement files. Ray “Pop” Marquez is documented to have negotiated his granddaughter Marisol into the slow cohort, at some cost to himself; the available record does not reveal on what basis a surface-handler is permitted to negotiate placement, and Pop appears to have known. An accelerated cohort exists at four-year periodicity for what the apparatus calls “high-load substrates” — the term is not defined in the public record, and three plausible interpretations (high-cognitive-uplift subjects, surface-handler-adjacent subjects, hybrid-class subjects with unstable satiation profiles) are all consistent with what can be triangulated from the abductee literature.

With a substrate registry on the order of several hundred million subjects and a standard cycle of seven years, the apparatus is running, somewhere on the order of, a hundred thousand maintenance procedures per day, every day, world over, and has been doing so since before the spec was written. That is the scale at which we are maintained.

Closing

An account that proposes humans-as-experimental-population without specifying the maintenance procedure is load-bearing without a load path. The substrate-claim only does work if there is an operation by which the experimental population is observed, recalibrated, and held in the epistemic condition the study requires. The maintenance procedure is that operation. It is not the only one — the operators also work us through climate-scale and culture-scale interventions later chapters will address — but it is the operation closest to the subject, performed on the subject’s own anatomy, every seven years, by technicians who have read the file.

The cribriform is the road. The calcite is the listening post. The owl is the lock on the door, and the door has been locked, very gently and very thoroughly, for six thousand years. We are observed. We have always been observed. The procedure is not the worst thing that could be true about us. The worst thing that could be true about us is that we were observed by no one at all — that the species had been allowed to come up on the surface, in this corner of the galaxy, with no older intelligence taking the trouble to notice.

I happen to think we are lucky. Lucky that the binding-image, when an older intelligence had to pick one and live with it for millennia, turned out to be something as benign as an owl. Lucky to be an experimental population that has, on the available evidence, been treated with relative care by the researchers running the study.